Compressed linear algebra for large-scale machine learning View Full Text


Ontology type: schema:ScholarlyArticle     


Article Info

DATE

2018-10

AUTHORS

Ahmed Elgohary, Matthias Boehm, Peter J. Haas, Frederick R. Reiss, Berthold Reinwald

ABSTRACT

Large-scale machine learning algorithms are often iterative, using repeated read-only data access and I/O-bound matrix-vector multiplications to converge to an optimal model. It is crucial for performance to fit the data into single-node or distributed main memory and enable fast matrix-vector operations on in-memory data. General-purpose, heavy- and lightweight compression techniques struggle to achieve both good compression ratios and fast decompression speed to enable block-wise uncompressed operations. Therefore, we initiate work—inspired by database compression and sparse matrix formats—on value-based compressed linear algebra (CLA), in which heterogeneous, lightweight database compression techniques are applied to matrices, and then linear algebra operations such as matrix-vector multiplication are executed directly on the compressed representation. We contribute effective column compression schemes, cache-conscious operations, and an efficient sampling-based compression algorithm. Our experiments show that CLA achieves in-memory operations performance close to the uncompressed case and good compression ratios, which enables fitting substantially larger datasets into available memory. We thereby obtain significant end-to-end performance improvements up to 9.2x. More... »

PAGES

719-744

References to SciGraph publications

  • 2011. The Architecture of SciDB in SCIENTIFIC AND STATISTICAL DATABASE MANAGEMENT
  • 2015-05. Deep learning in NATURE
  • 2014-12. The Stratosphere platform for big data analytics in THE VLDB JOURNAL
  • 1999. Numerical Optimization in NONE
  • Identifiers

    URI

    http://scigraph.springernature.com/pub.10.1007/s00778-017-0478-1

    DOI

    http://dx.doi.org/10.1007/s00778-017-0478-1

    DIMENSIONS

    https://app.dimensions.ai/details/publication/pub.1091609152


    Indexing Status Check whether this publication has been indexed by Scopus and Web Of Science using the SN Indexing Status Tool
    Incoming Citations Browse incoming citations for this publication using opencitations.net

    JSON-LD is the canonical representation for SciGraph data.

    TIP: You can open this SciGraph record using an external JSON-LD service: JSON-LD Playground Google SDTT

    [
      {
        "@context": "https://springernature.github.io/scigraph/jsonld/sgcontext.json", 
        "about": [
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/0801", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Artificial Intelligence and Image Processing", 
            "type": "DefinedTerm"
          }, 
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/08", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Information and Computing Sciences", 
            "type": "DefinedTerm"
          }
        ], 
        "author": [
          {
            "affiliation": {
              "alternateName": "University of Maryland, College Park", 
              "id": "https://www.grid.ac/institutes/grid.164295.d", 
              "name": [
                "University of Maryland, College Park, MD, USA"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Elgohary", 
            "givenName": "Ahmed", 
            "id": "sg:person.016142002337.07", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.016142002337.07"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "IBM Research - Almaden", 
              "id": "https://www.grid.ac/institutes/grid.481551.c", 
              "name": [
                "IBM Research \u2013 Almaden, San Jose, CA, USA"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Boehm", 
            "givenName": "Matthias", 
            "id": "sg:person.011163776213.64", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.011163776213.64"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "IBM Research - Almaden", 
              "id": "https://www.grid.ac/institutes/grid.481551.c", 
              "name": [
                "IBM Research \u2013 Almaden, San Jose, CA, USA"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Haas", 
            "givenName": "Peter J.", 
            "id": "sg:person.012400206062.53", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.012400206062.53"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "IBM Research - Almaden", 
              "id": "https://www.grid.ac/institutes/grid.481551.c", 
              "name": [
                "IBM Research \u2013 Almaden, San Jose, CA, USA"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Reiss", 
            "givenName": "Frederick R.", 
            "id": "sg:person.014037465351.39", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.014037465351.39"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "IBM Research - Almaden", 
              "id": "https://www.grid.ac/institutes/grid.481551.c", 
              "name": [
                "IBM Research \u2013 Almaden, San Jose, CA, USA"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Reinwald", 
            "givenName": "Berthold", 
            "id": "sg:person.014615154537.97", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.014615154537.97"
            ], 
            "type": "Person"
          }
        ], 
        "citation": [
          {
            "id": "https://doi.org/10.1145/2723372.2751520", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1001200275"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2597652.2597678", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1002768830"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/1807167.1807275", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1005716063"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1038/nature14539", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1010020120", 
              "https://doi.org/10.1038/nature14539"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2939672.2939864", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1011722546"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2463676.2465273", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1012500106"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/b98874", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1015038119", 
              "https://doi.org/10.1007/b98874"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/b98874", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1015038119", 
              "https://doi.org/10.1007/b98874"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/b98874", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1015038119", 
              "https://doi.org/10.1007/b98874"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2939672.2939856", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1015129996"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/1993636.1993727", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1015729665"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/3003665.3003669", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1016227050"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2213836.2213855", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1017718701"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/1654059.1654078", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1020511480"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-22351-8_1", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1021322907", 
              "https://doi.org/10.1007/978-3-642-22351-8_1"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-22351-8_1", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1021322907", 
              "https://doi.org/10.1007/978-3-642-22351-8_1"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2882903.2882939", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1023178958"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2877204", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1026597557"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2020408.2020558", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1027741969"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2882903.2882925", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1029760789"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/335168.335230", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1031398154"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2723372.2723713", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1032088310"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/362084.362137", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1033276736"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2618243.2618268", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1034629606"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2618243.2618254", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1035172347"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/1132863.1132864", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1035984289"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2983323.2983735", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1036836015"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2723372.2723712", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1038088578"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/1341531.1341547", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1038823884"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/1559845.1559877", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1040461261"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/1989323.1989399", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1040670879"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/1989323.1989448", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1040853109"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2688500.2688521", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1042980354"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/1951365.1951420", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1042986763"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2723372.2749432", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1043288069"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/s00778-014-0357-y", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1045392144", 
              "https://doi.org/10.1007/s00778-014-0357-y"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/1366230.1366244", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1046164365"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/1327452.1327492", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1047364446"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/214956.214963", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1049624778"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1145/2939672.2939675", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1052772499"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1080/01621459.1998.10473807", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1058305446"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1093/biomet/40.3-4.237", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1059416337"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/tpds.2012.290", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1061754094"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/tse.1985.231852", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1061787765"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/tsp.2016.2516971", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1061805372"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1137/0203025", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1062841239"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1137/070701649", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1062851552"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1137/090771806", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1062856710"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14778/1687553.1687573", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1067367496"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14778/1687553.1687576", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1067367498"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14778/1687627.1687671", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1067367574"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14778/2021017.2021023", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1067367895"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14778/2535573.2488340", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1067368150"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14778/2536222.2536233", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1067368178"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14778/2824032.2824087", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1067368763"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14778/2994509.2994515", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1067368922"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14778/3007263.3007279", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1067368962"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14778/3007263.3007312", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1067368995"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/soac.1991.143840", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1086287365"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14778/3025111.3025117", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1087132754"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14778/3137628.3137633", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1091641266"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.14778/3137628.3137664", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1091641297"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/icde.2017.108", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1093225447"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/icde.2012.12", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1093314546"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/icde.2006.150", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1093460056"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/icde.2011.5767930", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1093900074"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/dcc.2001.917151", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1094013775"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/ccp.2011.41", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1094785481"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/icde.2010.5447871", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1094893745"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/msst.2012.6232381", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1095329056"
            ], 
            "type": "CreativeWork"
          }
        ], 
        "datePublished": "2018-10", 
        "datePublishedReg": "2018-10-01", 
        "description": "Large-scale machine learning algorithms are often iterative, using repeated read-only data access and I/O-bound matrix-vector multiplications to converge to an optimal model. It is crucial for performance to fit the data into single-node or distributed main memory and enable fast matrix-vector operations on in-memory data. General-purpose, heavy- and lightweight compression techniques struggle to achieve both good compression ratios and fast decompression speed to enable block-wise uncompressed operations. Therefore, we initiate work\u2014inspired by database compression and sparse matrix formats\u2014on value-based compressed linear algebra (CLA), in which heterogeneous, lightweight database compression techniques are applied to matrices, and then linear algebra operations such as matrix-vector multiplication are executed directly on the compressed representation. We contribute effective column compression schemes, cache-conscious operations, and an efficient sampling-based compression algorithm. Our experiments show that CLA achieves in-memory operations performance close to the uncompressed case and good compression ratios, which enables fitting substantially larger datasets into available memory. We thereby obtain significant end-to-end performance improvements up to 9.2x.", 
        "genre": "research_article", 
        "id": "sg:pub.10.1007/s00778-017-0478-1", 
        "inLanguage": [
          "en"
        ], 
        "isAccessibleForFree": false, 
        "isPartOf": [
          {
            "id": "sg:journal.1044889", 
            "issn": [
              "1066-8888", 
              "0949-877X"
            ], 
            "name": "The VLDB Journal", 
            "type": "Periodical"
          }, 
          {
            "issueNumber": "5", 
            "type": "PublicationIssue"
          }, 
          {
            "type": "PublicationVolume", 
            "volumeNumber": "27"
          }
        ], 
        "name": "Compressed linear algebra for large-scale machine learning", 
        "pagination": "719-744", 
        "productId": [
          {
            "name": "readcube_id", 
            "type": "PropertyValue", 
            "value": [
              "d2d98261e0a2445fc09adfc597851b5f63bb80020e4a1e31a0e98b01ef4b5696"
            ]
          }, 
          {
            "name": "doi", 
            "type": "PropertyValue", 
            "value": [
              "10.1007/s00778-017-0478-1"
            ]
          }, 
          {
            "name": "dimensions_id", 
            "type": "PropertyValue", 
            "value": [
              "pub.1091609152"
            ]
          }
        ], 
        "sameAs": [
          "https://doi.org/10.1007/s00778-017-0478-1", 
          "https://app.dimensions.ai/details/publication/pub.1091609152"
        ], 
        "sdDataset": "articles", 
        "sdDatePublished": "2019-04-11T02:17", 
        "sdLicense": "https://scigraph.springernature.com/explorer/license/", 
        "sdPublisher": {
          "name": "Springer Nature - SN SciGraph project", 
          "type": "Organization"
        }, 
        "sdSource": "s3://com-uberresearch-data-dimensions-target-20181106-alternative/cleanup/v134/2549eaecd7973599484d7c17b260dba0a4ecb94b/merge/v9/a6c9fde33151104705d4d7ff012ea9563521a3ce/jats-lookup/v90/0000000001_0000000264/records_8700_00000541.jsonl", 
        "type": "ScholarlyArticle", 
        "url": "https://link.springer.com/10.1007%2Fs00778-017-0478-1"
      }
    ]
     

    Download the RDF metadata as:  json-ld nt turtle xml License info

    HOW TO GET THIS DATA PROGRAMMATICALLY:

    JSON-LD is a popular format for linked data which is fully compatible with JSON.

    curl -H 'Accept: application/ld+json' 'https://scigraph.springernature.com/pub.10.1007/s00778-017-0478-1'

    N-Triples is a line-based linked data format ideal for batch operations.

    curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/pub.10.1007/s00778-017-0478-1'

    Turtle is a human-readable linked data format.

    curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/pub.10.1007/s00778-017-0478-1'

    RDF/XML is a standard XML format for linked data.

    curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/pub.10.1007/s00778-017-0478-1'


     

    This table displays all metadata directly associated to this object as RDF triples.

    297 TRIPLES      21 PREDICATES      94 URIs      19 LITERALS      7 BLANK NODES

    Subject Predicate Object
    1 sg:pub.10.1007/s00778-017-0478-1 schema:about anzsrc-for:08
    2 anzsrc-for:0801
    3 schema:author N1fe02b5d5bd2486b8f62fc01f8a8def1
    4 schema:citation sg:pub.10.1007/978-3-642-22351-8_1
    5 sg:pub.10.1007/b98874
    6 sg:pub.10.1007/s00778-014-0357-y
    7 sg:pub.10.1038/nature14539
    8 https://doi.org/10.1080/01621459.1998.10473807
    9 https://doi.org/10.1093/biomet/40.3-4.237
    10 https://doi.org/10.1109/ccp.2011.41
    11 https://doi.org/10.1109/dcc.2001.917151
    12 https://doi.org/10.1109/icde.2006.150
    13 https://doi.org/10.1109/icde.2010.5447871
    14 https://doi.org/10.1109/icde.2011.5767930
    15 https://doi.org/10.1109/icde.2012.12
    16 https://doi.org/10.1109/icde.2017.108
    17 https://doi.org/10.1109/msst.2012.6232381
    18 https://doi.org/10.1109/soac.1991.143840
    19 https://doi.org/10.1109/tpds.2012.290
    20 https://doi.org/10.1109/tse.1985.231852
    21 https://doi.org/10.1109/tsp.2016.2516971
    22 https://doi.org/10.1137/0203025
    23 https://doi.org/10.1137/070701649
    24 https://doi.org/10.1137/090771806
    25 https://doi.org/10.1145/1132863.1132864
    26 https://doi.org/10.1145/1327452.1327492
    27 https://doi.org/10.1145/1341531.1341547
    28 https://doi.org/10.1145/1366230.1366244
    29 https://doi.org/10.1145/1559845.1559877
    30 https://doi.org/10.1145/1654059.1654078
    31 https://doi.org/10.1145/1807167.1807275
    32 https://doi.org/10.1145/1951365.1951420
    33 https://doi.org/10.1145/1989323.1989399
    34 https://doi.org/10.1145/1989323.1989448
    35 https://doi.org/10.1145/1993636.1993727
    36 https://doi.org/10.1145/2020408.2020558
    37 https://doi.org/10.1145/214956.214963
    38 https://doi.org/10.1145/2213836.2213855
    39 https://doi.org/10.1145/2463676.2465273
    40 https://doi.org/10.1145/2597652.2597678
    41 https://doi.org/10.1145/2618243.2618254
    42 https://doi.org/10.1145/2618243.2618268
    43 https://doi.org/10.1145/2688500.2688521
    44 https://doi.org/10.1145/2723372.2723712
    45 https://doi.org/10.1145/2723372.2723713
    46 https://doi.org/10.1145/2723372.2749432
    47 https://doi.org/10.1145/2723372.2751520
    48 https://doi.org/10.1145/2877204
    49 https://doi.org/10.1145/2882903.2882925
    50 https://doi.org/10.1145/2882903.2882939
    51 https://doi.org/10.1145/2939672.2939675
    52 https://doi.org/10.1145/2939672.2939856
    53 https://doi.org/10.1145/2939672.2939864
    54 https://doi.org/10.1145/2983323.2983735
    55 https://doi.org/10.1145/3003665.3003669
    56 https://doi.org/10.1145/335168.335230
    57 https://doi.org/10.1145/362084.362137
    58 https://doi.org/10.14778/1687553.1687573
    59 https://doi.org/10.14778/1687553.1687576
    60 https://doi.org/10.14778/1687627.1687671
    61 https://doi.org/10.14778/2021017.2021023
    62 https://doi.org/10.14778/2535573.2488340
    63 https://doi.org/10.14778/2536222.2536233
    64 https://doi.org/10.14778/2824032.2824087
    65 https://doi.org/10.14778/2994509.2994515
    66 https://doi.org/10.14778/3007263.3007279
    67 https://doi.org/10.14778/3007263.3007312
    68 https://doi.org/10.14778/3025111.3025117
    69 https://doi.org/10.14778/3137628.3137633
    70 https://doi.org/10.14778/3137628.3137664
    71 schema:datePublished 2018-10
    72 schema:datePublishedReg 2018-10-01
    73 schema:description Large-scale machine learning algorithms are often iterative, using repeated read-only data access and I/O-bound matrix-vector multiplications to converge to an optimal model. It is crucial for performance to fit the data into single-node or distributed main memory and enable fast matrix-vector operations on in-memory data. General-purpose, heavy- and lightweight compression techniques struggle to achieve both good compression ratios and fast decompression speed to enable block-wise uncompressed operations. Therefore, we initiate work—inspired by database compression and sparse matrix formats—on value-based compressed linear algebra (CLA), in which heterogeneous, lightweight database compression techniques are applied to matrices, and then linear algebra operations such as matrix-vector multiplication are executed directly on the compressed representation. We contribute effective column compression schemes, cache-conscious operations, and an efficient sampling-based compression algorithm. Our experiments show that CLA achieves in-memory operations performance close to the uncompressed case and good compression ratios, which enables fitting substantially larger datasets into available memory. We thereby obtain significant end-to-end performance improvements up to 9.2x.
    74 schema:genre research_article
    75 schema:inLanguage en
    76 schema:isAccessibleForFree false
    77 schema:isPartOf N6e07984291d4474297120f44c399c263
    78 Naeebb5bb7e654f2d8066f3d58901bd44
    79 sg:journal.1044889
    80 schema:name Compressed linear algebra for large-scale machine learning
    81 schema:pagination 719-744
    82 schema:productId N301dd0698ec04cea8a12f0328245fea6
    83 N7b6948f445444bca8806ccda3dee475d
    84 N964f679e7e25425bb212179e9aafe03d
    85 schema:sameAs https://app.dimensions.ai/details/publication/pub.1091609152
    86 https://doi.org/10.1007/s00778-017-0478-1
    87 schema:sdDatePublished 2019-04-11T02:17
    88 schema:sdLicense https://scigraph.springernature.com/explorer/license/
    89 schema:sdPublisher Na1aaadd4e61d4fc28a16423430bd028f
    90 schema:url https://link.springer.com/10.1007%2Fs00778-017-0478-1
    91 sgo:license sg:explorer/license/
    92 sgo:sdDataset articles
    93 rdf:type schema:ScholarlyArticle
    94 N025e38d270754a9fb737ea061f490311 rdf:first sg:person.014615154537.97
    95 rdf:rest rdf:nil
    96 N1fe02b5d5bd2486b8f62fc01f8a8def1 rdf:first sg:person.016142002337.07
    97 rdf:rest N9a9b81c87e85468c975f91f951a905bb
    98 N301dd0698ec04cea8a12f0328245fea6 schema:name doi
    99 schema:value 10.1007/s00778-017-0478-1
    100 rdf:type schema:PropertyValue
    101 N621bcc44caae4668afa6eb7e5b7cbbc2 rdf:first sg:person.014037465351.39
    102 rdf:rest N025e38d270754a9fb737ea061f490311
    103 N6e07984291d4474297120f44c399c263 schema:issueNumber 5
    104 rdf:type schema:PublicationIssue
    105 N7b6948f445444bca8806ccda3dee475d schema:name readcube_id
    106 schema:value d2d98261e0a2445fc09adfc597851b5f63bb80020e4a1e31a0e98b01ef4b5696
    107 rdf:type schema:PropertyValue
    108 N964f679e7e25425bb212179e9aafe03d schema:name dimensions_id
    109 schema:value pub.1091609152
    110 rdf:type schema:PropertyValue
    111 N9a9b81c87e85468c975f91f951a905bb rdf:first sg:person.011163776213.64
    112 rdf:rest Nb322191aaed84f459541cd11f3008b32
    113 Na1aaadd4e61d4fc28a16423430bd028f schema:name Springer Nature - SN SciGraph project
    114 rdf:type schema:Organization
    115 Naeebb5bb7e654f2d8066f3d58901bd44 schema:volumeNumber 27
    116 rdf:type schema:PublicationVolume
    117 Nb322191aaed84f459541cd11f3008b32 rdf:first sg:person.012400206062.53
    118 rdf:rest N621bcc44caae4668afa6eb7e5b7cbbc2
    119 anzsrc-for:08 schema:inDefinedTermSet anzsrc-for:
    120 schema:name Information and Computing Sciences
    121 rdf:type schema:DefinedTerm
    122 anzsrc-for:0801 schema:inDefinedTermSet anzsrc-for:
    123 schema:name Artificial Intelligence and Image Processing
    124 rdf:type schema:DefinedTerm
    125 sg:journal.1044889 schema:issn 0949-877X
    126 1066-8888
    127 schema:name The VLDB Journal
    128 rdf:type schema:Periodical
    129 sg:person.011163776213.64 schema:affiliation https://www.grid.ac/institutes/grid.481551.c
    130 schema:familyName Boehm
    131 schema:givenName Matthias
    132 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.011163776213.64
    133 rdf:type schema:Person
    134 sg:person.012400206062.53 schema:affiliation https://www.grid.ac/institutes/grid.481551.c
    135 schema:familyName Haas
    136 schema:givenName Peter J.
    137 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.012400206062.53
    138 rdf:type schema:Person
    139 sg:person.014037465351.39 schema:affiliation https://www.grid.ac/institutes/grid.481551.c
    140 schema:familyName Reiss
    141 schema:givenName Frederick R.
    142 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.014037465351.39
    143 rdf:type schema:Person
    144 sg:person.014615154537.97 schema:affiliation https://www.grid.ac/institutes/grid.481551.c
    145 schema:familyName Reinwald
    146 schema:givenName Berthold
    147 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.014615154537.97
    148 rdf:type schema:Person
    149 sg:person.016142002337.07 schema:affiliation https://www.grid.ac/institutes/grid.164295.d
    150 schema:familyName Elgohary
    151 schema:givenName Ahmed
    152 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.016142002337.07
    153 rdf:type schema:Person
    154 sg:pub.10.1007/978-3-642-22351-8_1 schema:sameAs https://app.dimensions.ai/details/publication/pub.1021322907
    155 https://doi.org/10.1007/978-3-642-22351-8_1
    156 rdf:type schema:CreativeWork
    157 sg:pub.10.1007/b98874 schema:sameAs https://app.dimensions.ai/details/publication/pub.1015038119
    158 https://doi.org/10.1007/b98874
    159 rdf:type schema:CreativeWork
    160 sg:pub.10.1007/s00778-014-0357-y schema:sameAs https://app.dimensions.ai/details/publication/pub.1045392144
    161 https://doi.org/10.1007/s00778-014-0357-y
    162 rdf:type schema:CreativeWork
    163 sg:pub.10.1038/nature14539 schema:sameAs https://app.dimensions.ai/details/publication/pub.1010020120
    164 https://doi.org/10.1038/nature14539
    165 rdf:type schema:CreativeWork
    166 https://doi.org/10.1080/01621459.1998.10473807 schema:sameAs https://app.dimensions.ai/details/publication/pub.1058305446
    167 rdf:type schema:CreativeWork
    168 https://doi.org/10.1093/biomet/40.3-4.237 schema:sameAs https://app.dimensions.ai/details/publication/pub.1059416337
    169 rdf:type schema:CreativeWork
    170 https://doi.org/10.1109/ccp.2011.41 schema:sameAs https://app.dimensions.ai/details/publication/pub.1094785481
    171 rdf:type schema:CreativeWork
    172 https://doi.org/10.1109/dcc.2001.917151 schema:sameAs https://app.dimensions.ai/details/publication/pub.1094013775
    173 rdf:type schema:CreativeWork
    174 https://doi.org/10.1109/icde.2006.150 schema:sameAs https://app.dimensions.ai/details/publication/pub.1093460056
    175 rdf:type schema:CreativeWork
    176 https://doi.org/10.1109/icde.2010.5447871 schema:sameAs https://app.dimensions.ai/details/publication/pub.1094893745
    177 rdf:type schema:CreativeWork
    178 https://doi.org/10.1109/icde.2011.5767930 schema:sameAs https://app.dimensions.ai/details/publication/pub.1093900074
    179 rdf:type schema:CreativeWork
    180 https://doi.org/10.1109/icde.2012.12 schema:sameAs https://app.dimensions.ai/details/publication/pub.1093314546
    181 rdf:type schema:CreativeWork
    182 https://doi.org/10.1109/icde.2017.108 schema:sameAs https://app.dimensions.ai/details/publication/pub.1093225447
    183 rdf:type schema:CreativeWork
    184 https://doi.org/10.1109/msst.2012.6232381 schema:sameAs https://app.dimensions.ai/details/publication/pub.1095329056
    185 rdf:type schema:CreativeWork
    186 https://doi.org/10.1109/soac.1991.143840 schema:sameAs https://app.dimensions.ai/details/publication/pub.1086287365
    187 rdf:type schema:CreativeWork
    188 https://doi.org/10.1109/tpds.2012.290 schema:sameAs https://app.dimensions.ai/details/publication/pub.1061754094
    189 rdf:type schema:CreativeWork
    190 https://doi.org/10.1109/tse.1985.231852 schema:sameAs https://app.dimensions.ai/details/publication/pub.1061787765
    191 rdf:type schema:CreativeWork
    192 https://doi.org/10.1109/tsp.2016.2516971 schema:sameAs https://app.dimensions.ai/details/publication/pub.1061805372
    193 rdf:type schema:CreativeWork
    194 https://doi.org/10.1137/0203025 schema:sameAs https://app.dimensions.ai/details/publication/pub.1062841239
    195 rdf:type schema:CreativeWork
    196 https://doi.org/10.1137/070701649 schema:sameAs https://app.dimensions.ai/details/publication/pub.1062851552
    197 rdf:type schema:CreativeWork
    198 https://doi.org/10.1137/090771806 schema:sameAs https://app.dimensions.ai/details/publication/pub.1062856710
    199 rdf:type schema:CreativeWork
    200 https://doi.org/10.1145/1132863.1132864 schema:sameAs https://app.dimensions.ai/details/publication/pub.1035984289
    201 rdf:type schema:CreativeWork
    202 https://doi.org/10.1145/1327452.1327492 schema:sameAs https://app.dimensions.ai/details/publication/pub.1047364446
    203 rdf:type schema:CreativeWork
    204 https://doi.org/10.1145/1341531.1341547 schema:sameAs https://app.dimensions.ai/details/publication/pub.1038823884
    205 rdf:type schema:CreativeWork
    206 https://doi.org/10.1145/1366230.1366244 schema:sameAs https://app.dimensions.ai/details/publication/pub.1046164365
    207 rdf:type schema:CreativeWork
    208 https://doi.org/10.1145/1559845.1559877 schema:sameAs https://app.dimensions.ai/details/publication/pub.1040461261
    209 rdf:type schema:CreativeWork
    210 https://doi.org/10.1145/1654059.1654078 schema:sameAs https://app.dimensions.ai/details/publication/pub.1020511480
    211 rdf:type schema:CreativeWork
    212 https://doi.org/10.1145/1807167.1807275 schema:sameAs https://app.dimensions.ai/details/publication/pub.1005716063
    213 rdf:type schema:CreativeWork
    214 https://doi.org/10.1145/1951365.1951420 schema:sameAs https://app.dimensions.ai/details/publication/pub.1042986763
    215 rdf:type schema:CreativeWork
    216 https://doi.org/10.1145/1989323.1989399 schema:sameAs https://app.dimensions.ai/details/publication/pub.1040670879
    217 rdf:type schema:CreativeWork
    218 https://doi.org/10.1145/1989323.1989448 schema:sameAs https://app.dimensions.ai/details/publication/pub.1040853109
    219 rdf:type schema:CreativeWork
    220 https://doi.org/10.1145/1993636.1993727 schema:sameAs https://app.dimensions.ai/details/publication/pub.1015729665
    221 rdf:type schema:CreativeWork
    222 https://doi.org/10.1145/2020408.2020558 schema:sameAs https://app.dimensions.ai/details/publication/pub.1027741969
    223 rdf:type schema:CreativeWork
    224 https://doi.org/10.1145/214956.214963 schema:sameAs https://app.dimensions.ai/details/publication/pub.1049624778
    225 rdf:type schema:CreativeWork
    226 https://doi.org/10.1145/2213836.2213855 schema:sameAs https://app.dimensions.ai/details/publication/pub.1017718701
    227 rdf:type schema:CreativeWork
    228 https://doi.org/10.1145/2463676.2465273 schema:sameAs https://app.dimensions.ai/details/publication/pub.1012500106
    229 rdf:type schema:CreativeWork
    230 https://doi.org/10.1145/2597652.2597678 schema:sameAs https://app.dimensions.ai/details/publication/pub.1002768830
    231 rdf:type schema:CreativeWork
    232 https://doi.org/10.1145/2618243.2618254 schema:sameAs https://app.dimensions.ai/details/publication/pub.1035172347
    233 rdf:type schema:CreativeWork
    234 https://doi.org/10.1145/2618243.2618268 schema:sameAs https://app.dimensions.ai/details/publication/pub.1034629606
    235 rdf:type schema:CreativeWork
    236 https://doi.org/10.1145/2688500.2688521 schema:sameAs https://app.dimensions.ai/details/publication/pub.1042980354
    237 rdf:type schema:CreativeWork
    238 https://doi.org/10.1145/2723372.2723712 schema:sameAs https://app.dimensions.ai/details/publication/pub.1038088578
    239 rdf:type schema:CreativeWork
    240 https://doi.org/10.1145/2723372.2723713 schema:sameAs https://app.dimensions.ai/details/publication/pub.1032088310
    241 rdf:type schema:CreativeWork
    242 https://doi.org/10.1145/2723372.2749432 schema:sameAs https://app.dimensions.ai/details/publication/pub.1043288069
    243 rdf:type schema:CreativeWork
    244 https://doi.org/10.1145/2723372.2751520 schema:sameAs https://app.dimensions.ai/details/publication/pub.1001200275
    245 rdf:type schema:CreativeWork
    246 https://doi.org/10.1145/2877204 schema:sameAs https://app.dimensions.ai/details/publication/pub.1026597557
    247 rdf:type schema:CreativeWork
    248 https://doi.org/10.1145/2882903.2882925 schema:sameAs https://app.dimensions.ai/details/publication/pub.1029760789
    249 rdf:type schema:CreativeWork
    250 https://doi.org/10.1145/2882903.2882939 schema:sameAs https://app.dimensions.ai/details/publication/pub.1023178958
    251 rdf:type schema:CreativeWork
    252 https://doi.org/10.1145/2939672.2939675 schema:sameAs https://app.dimensions.ai/details/publication/pub.1052772499
    253 rdf:type schema:CreativeWork
    254 https://doi.org/10.1145/2939672.2939856 schema:sameAs https://app.dimensions.ai/details/publication/pub.1015129996
    255 rdf:type schema:CreativeWork
    256 https://doi.org/10.1145/2939672.2939864 schema:sameAs https://app.dimensions.ai/details/publication/pub.1011722546
    257 rdf:type schema:CreativeWork
    258 https://doi.org/10.1145/2983323.2983735 schema:sameAs https://app.dimensions.ai/details/publication/pub.1036836015
    259 rdf:type schema:CreativeWork
    260 https://doi.org/10.1145/3003665.3003669 schema:sameAs https://app.dimensions.ai/details/publication/pub.1016227050
    261 rdf:type schema:CreativeWork
    262 https://doi.org/10.1145/335168.335230 schema:sameAs https://app.dimensions.ai/details/publication/pub.1031398154
    263 rdf:type schema:CreativeWork
    264 https://doi.org/10.1145/362084.362137 schema:sameAs https://app.dimensions.ai/details/publication/pub.1033276736
    265 rdf:type schema:CreativeWork
    266 https://doi.org/10.14778/1687553.1687573 schema:sameAs https://app.dimensions.ai/details/publication/pub.1067367496
    267 rdf:type schema:CreativeWork
    268 https://doi.org/10.14778/1687553.1687576 schema:sameAs https://app.dimensions.ai/details/publication/pub.1067367498
    269 rdf:type schema:CreativeWork
    270 https://doi.org/10.14778/1687627.1687671 schema:sameAs https://app.dimensions.ai/details/publication/pub.1067367574
    271 rdf:type schema:CreativeWork
    272 https://doi.org/10.14778/2021017.2021023 schema:sameAs https://app.dimensions.ai/details/publication/pub.1067367895
    273 rdf:type schema:CreativeWork
    274 https://doi.org/10.14778/2535573.2488340 schema:sameAs https://app.dimensions.ai/details/publication/pub.1067368150
    275 rdf:type schema:CreativeWork
    276 https://doi.org/10.14778/2536222.2536233 schema:sameAs https://app.dimensions.ai/details/publication/pub.1067368178
    277 rdf:type schema:CreativeWork
    278 https://doi.org/10.14778/2824032.2824087 schema:sameAs https://app.dimensions.ai/details/publication/pub.1067368763
    279 rdf:type schema:CreativeWork
    280 https://doi.org/10.14778/2994509.2994515 schema:sameAs https://app.dimensions.ai/details/publication/pub.1067368922
    281 rdf:type schema:CreativeWork
    282 https://doi.org/10.14778/3007263.3007279 schema:sameAs https://app.dimensions.ai/details/publication/pub.1067368962
    283 rdf:type schema:CreativeWork
    284 https://doi.org/10.14778/3007263.3007312 schema:sameAs https://app.dimensions.ai/details/publication/pub.1067368995
    285 rdf:type schema:CreativeWork
    286 https://doi.org/10.14778/3025111.3025117 schema:sameAs https://app.dimensions.ai/details/publication/pub.1087132754
    287 rdf:type schema:CreativeWork
    288 https://doi.org/10.14778/3137628.3137633 schema:sameAs https://app.dimensions.ai/details/publication/pub.1091641266
    289 rdf:type schema:CreativeWork
    290 https://doi.org/10.14778/3137628.3137664 schema:sameAs https://app.dimensions.ai/details/publication/pub.1091641297
    291 rdf:type schema:CreativeWork
    292 https://www.grid.ac/institutes/grid.164295.d schema:alternateName University of Maryland, College Park
    293 schema:name University of Maryland, College Park, MD, USA
    294 rdf:type schema:Organization
    295 https://www.grid.ac/institutes/grid.481551.c schema:alternateName IBM Research - Almaden
    296 schema:name IBM Research – Almaden, San Jose, CA, USA
    297 rdf:type schema:Organization
     




    Preview window. Press ESC to close (or click here)


    ...